Datasets
Models
Downloads: 1
| # | Question Report | Unnamed: 1 | Unnamed: 2 | Unnamed: 3 | Unnamed: 4 | Unnamed: 5 | Unnamed: 6 | Unnamed: 7 | Unnamed: 8 |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Report Generated: | 2023-04-18 16:36:00 | nan | nan | nan | nan | nan | nan | nan |
| 2 | Topic | Webinar ID | Actual Start Time | Actual Duration (minutes) | # Question | nan | nan | nan | nan |
| 3 | ASA-SSGG Short Course Series: Selective Introduction to Multi-Omics Analysis | 950 5160 1391 | 2023-04-18 14:30:00 | 121 | 25 | nan | nan | nan | nan |
| 4 | Question Details | nan | nan | nan | nan | nan | nan | nan | nan |
| 5 | # | Question | Asker Name | Asker Email | Answer | Question Time | Answered Time | Answer Name | Answer Email |
| 6 | 1 | how to download the whole things in lecure 3 folder...? | dhe2 | '- | 1) go to GitHub https://github.com/KechrisLab/ASAShortCourse-MultiOmics 2) click the “code” and select download ZIP | 2023-04-18 15:02:14 | 2023-04-18 15:04:37 | Rick Chang | rick.chang@pitt.edu |
| 7 | 1 | how to download the whole things in lecure 3 folder...? | dhe2 | '- | or type this command on terminal git clone https://github.com/KechrisLab/ASAShortCourse-MultiOmics.git | 2023-04-18 15:02:14 | 2023-04-18 15:05:48 | Rick Chang | rick.chang@pitt.edu |
| 8 | 1 | how to download the whole things in lecure 3 folder...? | dhe2 | '- | What Rick provided can download everything of the short course. For downloading a specific folder (e.g., lecture 3), I also cannot find a place to do it. | 2023-04-18 15:02:14 | 2023-04-18 15:08:53 | George Tseng | ctseng@pitt.edu |
| 9 | 2 | I just copied and pasted my questions and Jack’s answer. Q: .html has no library calls at the top. We need to add this lines of codes library("r.jive") library("TCGAbiolinks") library("survival") library("ggplot2") library("ggfortify") A: Thank you for pointing that out - it looks like the code folding did not work for the library calls, for some reason. Hopefully it is clear to attendees from the .Rmd file. | Hyunkyu Lee | '- | nan | 2023-04-18 15:02:35 | nan | nan | nan |
| 10 | 3 | Hi there - I was unable to install the TCGAbiolinks package - is there a trick to this one? | Anonymous Attendee | nan | What error message are you getting? | 2023-04-18 15:02:39 | 2023-04-18 15:03:44 | Sierra Niemiec | sierra.niemiec@cuanschutz.edu |
| 11 | 4 | I tried copy link as, but what I got was some weird stuff | dhe2 | '- | nan | 2023-04-18 15:02:50 | nan | nan | nan |
| 12 | 5 | Warning message: package ‘TCGAbiolinks’ is not available for this version of R | Anonymous Attendee | nan | What’s your R version? | 2023-04-18 15:04:09 | 2023-04-18 15:04:32 | Sierra Niemiec | sierra.niemiec@cuanschutz.edu |
| 13 | 6 | ^have you used BiocManager::install("TCGAbiolinks") | Jenea I. Adams | '- | nan | 2023-04-18 15:04:17 | nan | nan | nan |
| 14 | 7 | 4.2.3 | Anonymous Attendee | nan | I think Jenea’s suggestion is good. Try: BiocManager::install("TCGAbiolinks") | 2023-04-18 15:04:37 | 2023-04-18 15:05:06 | Sierra Niemiec | sierra.niemiec@cuanschutz.edu |
| 15 | 8 | Great - it seems to be working! Thank you! | Anonymous Attendee | nan | Great! | 2023-04-18 15:06:36 | 2023-04-18 15:06:58 | Sierra Niemiec | sierra.niemiec@cuanschutz.edu |
| 16 | 9 | to Rick Chang: thank you, it now works! | dhe2 | '- | nan | 2023-04-18 15:07:35 | nan | nan | nan |
| 17 | 10 | can you use pca as outliers detection for gene expression data? | Anna Giczewska | '- | Yes, PCA as well as advanced tSNE and UMAP methods have been used for outlier detection and batch effect detection. | 2023-04-18 15:16:37 | 2023-04-18 15:18:09 | George Tseng | ctseng@pitt.edu |
| 18 | 11 | thank you | Anna Giczewska | '- | nan | 2023-04-18 15:18:23 | nan | nan | nan |
| 19 | 12 | yes please, share the link to the code that was refernce on slide 11 | Anna Giczewska | '- | Slide numbered 11 had text, are you referring to a different slide? Most of the presentation figures are from other references but we can ask Jack for any code that was used. | 2023-04-18 15:19:41 | 2023-04-18 15:22:40 | Nancy Zhang | asa.ssgg2018@gmail.com |
| 20 | 13 | Is there an instance where using PCA to discretize continuous multiomic data for downstream use would not be appropriate? | Jenea I. Adams | '- | If I understood correctly, I believe that you are asking about clustering samples into discrete classes through PCA or other methods? If so, there are data applications where clustering may not be appropriate, i.e., samples come from a continuum and it may not be appropriate to discretize them. Please see materials for Lecture 2, it discusses various metrics that can be used to assess clustering quality. Let me know if I understood your question correctly. | 2023-04-18 15:21:41 | 2023-04-18 15:27:45 | Nancy Zhang | asa.ssgg2018@gmail.com |
| 21 | 14 | Are there any integration method that would allow us to integrate multiomic data with different subsets of samples in a cohort? | Joanne TY Lim | '- | Most methods require matched data. But some of the clustering methods discused last week in Lecture 2 can handle unmatched data (e.g., NEMO). | 2023-04-18 15:25:13 | 2023-04-18 15:35:26 | Nancy Zhang | asa.ssgg2018@gmail.com |
| 22 | 15 | yes, that’s essentially what I asked, thank you | Jenea I. Adams | '- | nan | 2023-04-18 15:30:32 | nan | nan | nan |
| 23 | 16 | do intNMF and iCluster also use k-means consensus clustering? do certain types of clustering algorithm pair better with certain dimension reduction techniques? this question may not even make sense… | Timothy Vigers | '- | Clustering and dimension reduction are related problems but have different focus. Indeed, clustering based on dimension reduced data is a popular approach and iCluster is a good example (applying K-means after factor analysis). I have not seen evaluation of pairing different clustering methods and dimension reduction methods. Hope this answers your question. | 2023-04-18 15:33:32 | 2023-04-18 15:43:16 | George Tseng | ctseng@pitt.edu |
| 24 | 17 | How to interpretate the plots in p27? | Yanghyeon Cho | '- | Dot plots of the -log10 of the p-values for ten dimension reduced factors associated with survival in a Cox regression, stratified by method and cancer type. | 2023-04-18 15:42:02 | 2023-04-18 15:47:18 | Jack Pattee | jack.pattee@cuanschutz.edu |
| 25 | 18 | i might be seeing things, but it looks like the miRNA individual and noise heatmaps have very similar structure. any idea why that is (if it is the case)? | Timothy Vigers | '- | I agree, they do look similar in the heatmap but it’s hard to confirm based on the figure because of the high dimensionality. Looking into the additional information JIVE returns (through summary) and downstream analyses would be important for following up. | 2023-04-18 15:59:00 | 2023-04-18 16:14:27 | Nancy Zhang | asa.ssgg2018@gmail.com |
| 26 | 18 | i might be seeing things, but it looks like the miRNA individual and noise heatmaps have very similar structure. any idea why that is (if it is the case)? | Timothy Vigers | '- | Also, that structure seems similar due to a few miRNA. It would be good to check back on those features to see if there are any issues (low expression, small count, outliers, etc). | 2023-04-18 15:59:00 | 2023-04-18 16:16:02 | Nancy Zhang | asa.ssgg2018@gmail.com |
| 27 | 19 | Thanks so much - excellent talk and lab! | Jennifer Bryant | '- | nan | 2023-04-18 16:17:16 | nan | nan | nan |
| 28 | 20 | One question - in terms of sample size desired for the multi-omics analysis, would there be sort of guidance or recommendations? e.g., if I would have 50-100 patients with different layers of the molecular data, would such analysis still be reliable? Thank you! | Jun Wang | '- | Mathematically it should be okay to perform the dimension reduction (I think), but whether that dimension reduction is ‘good for anything’ might depend more on the sample size. | 2023-04-18 16:19:02 | 2023-04-18 16:21:22 | Jack Pattee | jack.pattee@cuanschutz.edu |
| 29 | 21 | Have you tried applying a spatial correlation measure to the heatmaps? It's hard to tell by eye whether there is any structure. | Monnie McGee | '- | Following up on our live answer, I’m not familiar with spatial approaches in this area but I looked online and found some literature on ‘heat map comparisons metrics’. But I’m not sure that has been included in this pacakge. | 2023-04-18 16:19:36 | 2023-04-18 16:24:18 | Nancy Zhang | asa.ssgg2018@gmail.com |
| 30 | 22 | Can you perform JIVE with regularization so that each ‘metagene’ only includes N or fewer number of features? | Hossein Asgharian | '- | live answered | 2023-04-18 16:19:50 | 2023-04-18 16:21:32 | Jack Pattee | jack.pattee@cuanschutz.edu |
| 31 | 23 | Patients data from clinical trials - post hoc analysis | Jun Wang | '- | I think this should be doable. | 2023-04-18 16:20:30 | 2023-04-18 16:20:48 | Jack Pattee | jack.pattee@cuanschutz.edu |
| 32 | 24 | Thank you for the comments and suggestion! | Jun Wang | '- | nan | 2023-04-18 16:21:24 | nan | nan | nan |
| 33 | 25 | Thanks for the great talk. My qustion is that: Based on the data given, is there a reasonable way to select the right (statistical) number of PCs? It seems JIVE also requires the parameters for the ideal number of PC or factors. isn't it? | Hyunkyu Lee | '- | There are statistical and visual methods for selecting the ‘optimal’ number of PCs used in reconstruction; often a scree plot or similar is used unless a higher level of mathematical precision is required. JIVE will automatically select the number of ‘PCs’ used in the reconstruction with a permutation-based or model-based criteria, as there is no easy visual way of doing this in a multi-dimensional context. | 2023-04-18 16:22:15 | 2023-04-18 16:23:38 | Jack Pattee | jack.pattee@cuanschutz.edu |